click to resample

Bio

I am Juil Koo, a fourth-year PhD student in the KAIST Visual AI Group advised by Prof. Minhyuk Sung, and a research intern at Meta Reality Labs Research. Previously, I interned at Adobe Research in 2024.

My research goal is to build embodied AI systems: agents that accomplish tasks in the physical world. I view this as a closed loop between an agent and its surrounding environment. My work addresses both sides of the loop through probabilistic generative modeling, learning distributions over visual worlds and agent behaviors.
In more concrete terms, I am interested in vision-language-action (VLA) and world(-action) models.

My Erdős number is 3 (Juil Koo → Leonidas Guibas → Frances Yao → Paul Erdős).

News

Publications

* denotes equal contribution.

Drifting Field Policy: A One-Step Generative Policy via Wasserstein Gradient Flow
Juil Koo, Mingue Park, Jiwon Choi, Yunhong Min, Minhyuk Sung NeurIPS 2026 [Project] [Paper] [Code] [Slides]
Token Warping Helps MLLMs Look from Nearby Viewpoints
Phillip Y. Lee*, Chanho Park*, Mingue Park, Seugwoo Yoo, Juil Koo, Minhyuk Sung CVPR 2026 [Project] [Paper] [Code]
Toward Ambulatory Vision: Learning Visually-Grounded Active View Selection
Juil Koo*, Daehyeon Choi*, Sangwoo Youn*, Phillip Y. Lee, Minhyuk Sung ICML 2026 Workshop RLxF [Project] [Paper] [Code]
BézierFlow: Bézier Stochastic Interpolant Schedulers for Few-Step Generation
Yunhong Min*, Juil Koo*, Seungwoo Yoo, Minhyuk Sung ICLR 2026 [Project] [Paper] [Code]
DiffusionRollout: Uncertainty-Aware Rollout Planning in Long-Horizon PDE Solving
Seungwoo Yoo*, Juil Koo*, Daehyeon Choi*, Minhyuk Sung TMLR [Paper]
BoxSplitGen: A Generative Model for 3D Part Bounding Boxes in Varying Granularity
Juil Koo*, Wei-Tung Lin*, Chanho Park, Chanhyeok Park, Minhyuk Sung WACV 2026 [Project] [Paper]
VideoHandles: Editing 3D Object Compositions in Videos Using Video Generative Priors
Juil Koo, Paul Guerrero, Chun-Hao Paul Huang, Duygu Ceylan, Minhyuk Sung CVPR 2025 [Project] [Paper]
SyncTweedies: A General Generative Framework Based on Synchronized Diffusions
Jaihoon Kim*, Juil Koo*, Kyeongmin Yeo*, Minhyuk Sung NeurIPS 2024 [Project] [Paper] [Code]
Neural Pose Representation Learning for Generating and Transferring Non-Rigid Object Poses
Seungwoo Yoo, Juil Koo, Kyeongmin Yeo, Minhyuk Sung NeurIPS 2024 [Project] [Paper] [Code]
Posterior Distillation Sampling
Juil Koo, Chanho Park, Minhyuk Sung CVPR 2024 [Project] [Paper] [Code] [Video] [PDF]
SALAD: Part-Level Latent Diffusion for 3D Shape Generation and Manipulation
Juil Koo*, Seungwoo Yoo*, Minh Hieu Nguyen*, Minhyuk Sung ICCV 2023 [Project] [Paper] [Code] [PDF] [Hugging Face]
PartGlot: Learning Shape Part Segmentation from Language Reference Games
Juil Koo, Ian Huang, Panos Achlioptas, Leonidas Guibas, Minhyuk Sung CVPR 2022 (Oral) [Project] [Paper] [Code] [Video]

Academic Service